Skip to contentSkip to conversation

xen-pciback: also support disabling of bus-mastering and memory-write-invalidate

PATCH Applied Jan Beulich 3 participants 11 messages Mar 11, 2015 – Mar 16, 2015 Xen hypervisor interface

lore
Share & tools
lore
#0 lore raw

PATCH 2/2 xen-pciback: also support disabling of bus-mastering and memory-write-invalidate

It's not clear to me why only the enabling operation got handled so far.

---
 drivers/xen/xen-pciback/conf_space_header.c |   16 +++++++++++++++-
 1 file changed, 15 insertions(+), 1 deletion(-)
drivers/xen/xen-pciback/conf_space_header.c +14 -1
@@ -88,9 +88,15 @@ static int command_write(struct pci_dev 
 			printk(KERN_DEBUG DRV_NAME ": %s: set bus master\n",
 			       pci_name(dev));
 		pci_set_master(dev);
+	} else if (dev->is_busmaster && !is_master_cmd(value)) {
+		if (unlikely(verbose_request))
+			printk(KERN_DEBUG DRV_NAME ": %s: clear bus master\n",
+			       pci_name(dev));
+		pci_clear_master(dev);
 	}
 
-	if (value & PCI_COMMAND_INVALIDATE) {
+	if (!(cmd->val & PCI_COMMAND_INVALIDATE) &&
+	    (value & PCI_COMMAND_INVALIDATE)) {
 		if (unlikely(verbose_request))
 			printk(KERN_DEBUG
 			       DRV_NAME ": %s: enable memory-write-invalidate\n",
@@ -101,6 +107,13 @@ static int command_write(struct pci_dev 
 				pci_name(dev), err);
 			value &= ~PCI_COMMAND_INVALIDATE;
 		}
+	} else if ((cmd->val & PCI_COMMAND_INVALIDATE) &&
+		   !(value & PCI_COMMAND_INVALIDATE)) {
+		if (unlikely(verbose_request))
+			printk(KERN_DEBUG
+			       DRV_NAME ": %s: disable memory-write-invalidate\n",
+			       pci_name(dev));
+		pci_clear_mwi(dev);
 	}
 
 	cmd->val = value;
Reviewed-by on v1 2/2
#1 lore raw
Quoted 2 lines · Jan Beulich

It's not clear to me why only the enabling operation got handled so far.

Quoted 42 lines
---
 drivers/xen/xen-pciback/conf_space_header.c |   16 +++++++++++++++-
 1 file changed, 15 insertions(+), 1 deletion(-)
drivers/xen/xen-pciback/conf_space_header.c +14 -1
@@ -88,9 +88,15 @@ static int command_write(struct pci_dev 
 			printk(KERN_DEBUG DRV_NAME ": %s: set bus master\n",
 			       pci_name(dev));
 		pci_set_master(dev);
+	} else if (dev->is_busmaster && !is_master_cmd(value)) {
+		if (unlikely(verbose_request))
+			printk(KERN_DEBUG DRV_NAME ": %s: clear bus master\n",
+			       pci_name(dev));
+		pci_clear_master(dev);
 	}
 
-	if (value & PCI_COMMAND_INVALIDATE) {
+	if (!(cmd->val & PCI_COMMAND_INVALIDATE) &&
+	    (value & PCI_COMMAND_INVALIDATE)) {
 		if (unlikely(verbose_request))
 			printk(KERN_DEBUG
 			       DRV_NAME ": %s: enable memory-write-invalidate\n",
@@ -101,6 +107,13 @@ static int command_write(struct pci_dev 
 				pci_name(dev), err);
 			value &= ~PCI_COMMAND_INVALIDATE;
 		}
+	} else if ((cmd->val & PCI_COMMAND_INVALIDATE) &&
+		   !(value & PCI_COMMAND_INVALIDATE)) {
+		if (unlikely(verbose_request))
+			printk(KERN_DEBUG
+			       DRV_NAME ": %s: disable memory-write-invalidate\n",
+			       pci_name(dev));
+		pci_clear_mwi(dev);
 	}
 
 	cmd->val = value;
Reviewed-by on v1 2/2
#2 lore raw
In reply to Jan Beulich
Quoted 2 lines · Jan Beulich

It's not clear to me why only the enabling operation got handled so far.

Quoted 42 lines
---
 drivers/xen/xen-pciback/conf_space_header.c |   16 +++++++++++++++-
 1 file changed, 15 insertions(+), 1 deletion(-)
drivers/xen/xen-pciback/conf_space_header.c +14 -1
@@ -88,9 +88,15 @@ static int command_write(struct pci_dev 
 			printk(KERN_DEBUG DRV_NAME ": %s: set bus master\n",
 			       pci_name(dev));
 		pci_set_master(dev);
+	} else if (dev->is_busmaster && !is_master_cmd(value)) {
+		if (unlikely(verbose_request))
+			printk(KERN_DEBUG DRV_NAME ": %s: clear bus master\n",
+			       pci_name(dev));
+		pci_clear_master(dev);
 	}
 
-	if (value & PCI_COMMAND_INVALIDATE) {
+	if (!(cmd->val & PCI_COMMAND_INVALIDATE) &&
+	    (value & PCI_COMMAND_INVALIDATE)) {
 		if (unlikely(verbose_request))
 			printk(KERN_DEBUG
 			       DRV_NAME ": %s: enable memory-write-invalidate\n",
@@ -101,6 +107,13 @@ static int command_write(struct pci_dev 
 				pci_name(dev), err);
 			value &= ~PCI_COMMAND_INVALIDATE;
 		}
+	} else if ((cmd->val & PCI_COMMAND_INVALIDATE) &&
+		   !(value & PCI_COMMAND_INVALIDATE)) {
+		if (unlikely(verbose_request))
+			printk(KERN_DEBUG
+			       DRV_NAME ": %s: disable memory-write-invalidate\n",
+			       pci_name(dev));
+		pci_clear_mwi(dev);
 	}
 
 	cmd->val = value;
#3 lore raw
Quoted 5 lines · Konrad Rzeszutek Wilk
Quoted 2 lines · Jan Beulich

It's not clear to me why only the enabling operation got handled so far.

With no reports that this actually fixes anything, do we want this in 4.0 or 4.1?

David

#5 lore raw
Quoted 9 lines · David Vrabel
Quoted 5 lines · Konrad Rzeszutek Wilk
Quoted 2 lines · Jan Beulich

It's not clear to me why only the enabling operation got handled so far.

With no reports that this actually fixes anything, do we want this in 4.0 or 4.1?

4.1 is fine.

Quoted 2 lines

David

Applied
#9 lore raw
In reply to Jan Beulich
Quoted 2 lines · Jan Beulich

It's not clear to me why only the enabling operation got handled so far.

Applied to devel/for-linus-4.1, thanks.

David

Applied
#10 lore raw
In reply to Jan Beulich
Quoted 2 lines · Jan Beulich

It's not clear to me why only the enabling operation got handled so far.

Applied to devel/for-linus-4.1, thanks.

David